This paper reports significant improvements in the efficacy of sequence-independent amplification and quality of sequencing of viruses with segmented double-stranded RNA (dsRNA) genomes. We demonstrate that most remaining bottlenecks in dsRNA virus genome characterization have now been eliminated. Both the amplification and sequencing technologies used require no previous sequence knowledge of the viral dsRNA, there is no longer a need to separate genome segments or amplicons and the sequence-determined bias observed in cloning has been overcome. Combining very efficient genome amplification with pyrophosphate-based 454 (GS20/ FLX) sequencing enabled sequencing of complete segmented dsRNA genomes and accelerated the sequence analysis of the amplified viral genomes. We report the complete consensus sequence of seven viruses from four different dsRNA virus groups, which include the first complete sequence of the genome of equine encephalosis virus (EEV), the first complete sequence of an African horsesickness virus (AHSV) genome determined directly from a blood sample and a complete human rotavirus genome determined from faeces. We also present the first comparison between the complete consensus sequence of a virulent and an attenuated strain of AHSV1. Ultra-deep sequencing (.400-fold coverage) of the AHSV1 reference and attenuated strains revealed different ratios of reassortants in the reference strain and allowed quasispecies detection in the plaque-purified attenuated strain of AHSV1. This approach amounts to a paradigm shift in dsRNA virus research, since it is sensitive and specific enough for comprehensive investigations of the evolution and genetic diversity in dsRNA virus populations.

Heartwater, a tick-borne disease of domestic and wild ruminants, is caused by the intracellular rickettsia Ehrlichia ruminantium (previously known as Cowdria ruminantium). It is a major constraint to livestock production throughout subSaharan Africa, and it threatens to invade the Americas, yet there is no immediate prospect of an effective vaccine. A shotgun genome sequencing project was undertaken in the expectation that access to the complete protein coding repertoire of the organism will facilitate the search for vaccine candidate genes. We report here the complete 1,516,355-bp sequence of the type strain, the stock derived from the South African Welgevonden isolate. Only 62% of the genome is predicted to be coding sequence, encoding 888 proteins and 41 stable RNA species. The most striking feature is the large number of tandemly repeated and duplicated sequences, some of continuously variable copy number, which contributes to the low proportion of coding sequence. These repeats have mediated numerous translocation and inversion events that have resulted in the duplication and truncation of some genes and have also given rise to new genes. There are 32 predicted pseudogenes, most of which are truncated fragments of genes associated with repeats. Rather then being the result of the reductive evolution seen in other intracellular bacteria, these pseudogenes appear to be the product of ongoing sequence duplication events.gene duplication ͉ bacterial genome ͉ molecular sequence data ͉ intracellular adaptation E hrlichia ruminantium (previously known as Cowdria ruminantium) is an obligate intracellular bacterium in the order Rickettsiales. Species in this order cause serious diseases in man and domestic animals throughout the world. E. ruminantium is transmitted by ticks of the genus Amblyomma and causes heartwater, a fatal and economically important disease of wild and domestic ruminants. The disease occurs throughout subSaharan Africa and on several Caribbean islands, from which it threatens to invade the Americas (1), but the existing immunization procedures are rudimentary and relatively ineffective (2). E. ruminantium is a fragile bacterium with exacting culture requirements in eukaryotic cell lines; genetic manipulation has not been attempted, and little is known about its mechanisms of virulence or pathogenesis. Heartwater affects all domestic ruminants, and 80-95% of naïve animals die within 3 weeks, but those that recover have a T cell-mediated immunity to subsequent homologous challenge (3). In the absence of any directed strategy to identify T cell-stimulatory proteins we sequenced the E. ruminantium genome in the expectation that access to the complete protein-coding repertoire of the organism would facilitate the search for vaccine candidate genes.